Re: Actual IPv6 deployment observed

[Danny McPherson <danny at tcb.net>]

On Jul 27, 2009, at 3:02 AM, Iljitsch van Beijnum wrote:

> On 27 jul 2009, at 9:43, Arnt Gulbrandsen wrote:
>
> > This must mean that silently enabling IPv6 increases the number of  
> > people for whom IPv6 works by a factor of around 100 (from <0.01%  
> > in the general population
>
> > (http://asert.arbornetworks.com/2008/08/the-end-is-near-but-is- 
> > ipv6/ said <0.01%.)
>
> The 0.01% they talk about is TRAFFIC, not USERS. And it's bogus  
> anyway.

Not that I want to have this discussion here again (folks should
revisit the archives), but it is not bogus, you just have to understand
what's being measured - and continues to be measured (and the same
measurement is today about 20x that .0025% from last August - I'll
write something up on this (with details again on methodology) within
the next month or two.

> For instance, just the traffic through the AMS-IX is many times more  
> than what they measured:

AMS-IX is an anomaly and not representative of the global Internet
in this regard, _IMO.  Then again, it is another data point, and
so long as folks know what's being measured, it is simply another
data point.

> http://www.ams-ix.net/technical/stats/sflow/
>
> (Note that they use meaningless lying graphs = don't start at 0.)
>
> At AMS-IX, native IPv6 traffic is now 0.3% on average, up from 0.1%  
> a year or so ago. When I did some web bug measurements years ago my  
> results where about 0.16% IPv6 users (1 in 666). Last year, Google  
> got 0.25%.

To your point, that Google stat was users,not traffic.  I suspect
their traffic rates are much lower than that :-)

-danny